Live Music Archive
Librivox Free Audio
Metropolitan Museum
Cleveland Museum of Art
Internet Arcade
Console Living Room
Books to Borrow
Open Library
TV News
Understanding 9/11Word2Vec Models for Twenty-year Periods of 18C (ECCO, "Literature and Language")
Item Preview
There Is No Preview Available For This Item
This item does not appear to have any files that can be experienced on Archive.org.
Show all files
Word2Vec Models for Twenty-year Periods of 18C (ECCO, "Literature and Language")
Creator: Ryan Heuser (heuser@stanford.edu | @quadrismegistus)
License: Creative Commons (CC BY-SA)
This is a zipped folder of five word2vec models. Each model was trained on 150 million words randomly sampled from the "Literature and Language" section of Eighteenth-Century Collections Online ("ECCO," by Gale). The five models correspond to the five twenty-year periods of the 18C: 1700-19, 1720-39, 1740-59, 1760-79, 1780-1799.
Each period has two files. One is a gzipped word2vec model, saved in Google's word2vec format, which is readable by word2vec, gensim, and GloVe. The other is a plain text file of the form [word][space][count][newline]. Both files can be easily imported with gensim like this:
import gensim
# Load word2vec model in the form: load_word2vec_format([gzipped model filename], [vocabulary filename])
model = gensim.models.Word2Vec.load_word2vec_format('word2vec.ECCO.1700-1719.skipgram_n=10.model.txt.gz', 'word2vec.ECCO.1700-1719.skipgram_n=10.model.vocab.txt')
# Test an analogy: Man is to woman as king is to ____?
print model.most_similar(['woman','king'], ['man'])
The models were trained using a skip-gram size of ten words. In order to reduce file size, each model was pruned of words not in the most frequent 50,000 words for that model / twenty-year period sample.
Reviews
There are no reviews yet. Be the first one to
write a review.
